IDENTIFICATION OF QTLs AND GENES FOR

DROUGHT TOLERANCE USING LINKAGE

MAPPING AND ASSOCIATION MAPPING

APPROACHES IN CHICKPEA (Cicer arietinum)

THESIS SUBMITTED TO

OSMANIA UNIVERSITY FOR THE AWARD OF

DOCTOR OF PHILOSOPHY IN GENETICS

Spurthi Nagesh Nayak

DEPARTMENT OF GENETICS OSMANIA UNIVERSITY, HYDERABAD

2010

CERTIFICATE

This is to certify that Ms. Spurthi Nagesh Nayak has carried out the research work

embodied in the present thesis entitled “Identification of QTLs and genes for drought

tolerance using linkage mapping and association mapping approaches in chickpea

(Cicer arietinum)” for the degree of Doctor of Philosophy under the joint-supervision of

Dr. Rajeev K Varshney, Principal Scientist, Applied Genomics Laboratory, International

Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru and Prof. P B

Kavi Kishor, Department of Genetics, Osmania University, Hyderabad.

This is an original research work carried out at ICRISAT and is satisfactory for the award

of Doctor of Philosophy. Any part of this work has not been submitted for the award of

any degree or diploma of any other University or Institute.

Dr. Rajeev K Varshney

Supervisor

Prof. P B Kavi Kishor

Co-supervisor

DECLARATION

I hereby declare that the research work presented in this thesis entitled “Identification of

QTLs and genes for drought tolerance using linkage mapping and association

mapping approaches in chickpea (Cicer arietinum)”, has been carried out under the

supervision of Dr. Rajeev K Varshney at International Crops Research Institute for Semi-

Arid tropics (ICRISAT), Patancheru and co-supervision of Prof. P B Kavi Kishor at

Department of Genetics, Osmania University, Hyderabad.

This work is original and no part of the thesis has been submitted earlier for the award of

any degree or diploma of any University.

Date: (Spurthi Nagesh Nayak)

Place: Hyderabad

ACKNOWLEDGMENTS This dissertation would not have been possible without the guidance and the help of several individuals who in one way or other contributed and extended their valuable assistance in the preparation and completion of this study.

First and foremost, I am sincerely obliged and indebted to my supervisor Dr. Rajeev K Varshney, Principal Scientist-Applied Genomics, ICRISAT and Leader Theme1-Comparative and Applied Genomics (CAG), Generation Challenge Programme for his peer supervision, constructive comments, enthusiastic discussions, endless support, encouragement and able direction throughout my research project. It was indeed a rare privilege for me to work under his able guidance. I also thank him from bottom of my heart for the critical evaluation and emending suggestions in the preparation of scientific papers for journals and in achieving the final form of dissertation.

I wish to record my profound gratitude and sincere thanks to the personification of generosity and kindness to my co-supervisor Prof. P B Kavi Kishor, Department of Genetics, Osmania University for his expedient advice, ever-encouraging suggestions, timely help, invaluable support, kind concern and consideration regarding my academic requirements during this tenure of research work.

I take this opportunity to thank Dr. Dave Hoisington for giving me opportunity to work in one of the best labs of International standards. With respect, regards and immense pleasure, I wish to acknowledge and express sincere thanks from my heart thanks to several scientists including Dr. Jayashree Balaji for the valuable suggestions during the sequence analysis; Dr. Junichi Kashiwagi and Dr. L Krishnamurthy for providing the phenotypic data; Dr. T Nepolean for his help during mapping; Dr. H D Upadhyaya and Dr. P M Gaur for providing the genetic material used in the present study.

It is my immense pleasure to render sincere regards to all the collaborators of the project especially Dr. Doug Cook (University of California, Davis, USA), Dr. Peter Winter (GenXPro/ University of Frankfurt, Germany), Dr. Dominique This (CIRAD, Montpellier, France) for their valuable support and co-operation to complete this invaluable piece of work. I also thank Dr. Pawan Kulwal (MPKV, Rahuri) and Dr. Joy K Roy (NABI, Mohali) for their help rendered during analysis. I take this opportunity to thank Dr. Sabhyata Bhatia (NIPGR, New Delhi) for kindly providing the microsatellite markers used in the present study. I render special thanks Dr. RR Mir, Visiting Scientist, ICRISAT for his help in finalizing the thesis.

With immense pleasure, I express my cordial thanks to Mrs. Seetha Kannan, Mrs. Manjula B, Mr. Prasad KDV and Mr. Murali Krishna Y for their administrative help. Words are less to express my gratitude to Mr. A Gafoor, Mr. K Eshwar and Mr. G Somaraju for teaching me the laboratory techniques at the beginning of the research work. I shall be failing in my duty if I do not express my cordial thanks to all my labmates and friends especially Srivani, Pavana, Nicy, Sivakumar, Mahendar, Ramu, Rajaram, Abhishek, Jalaja, Gautami, Susmitha-Krishna, Prathima, Ramesh, Anuja, Neha, Rachit, Ramani, Krithika and Sunita for their support during my happiness and hard times. I also thank all the friends and staff at Department of Genetics and the Principal Office of Osmania University, Hyderabad for their kind help and co-operation.

I also thank the Bioinformatics and Biometrics unit for the assistance given at the time of need. Assistance rendered by the members of Central Support Lab, library and Learning Systems Unit at ICRISAT is gratefully acknowledged.

I feel blessed to have my husband P Srinivasa Reddy in my life, who was the constant inspiration for me to carry out the research and gave emotional support whenever I had difficult times. I feel indebted to my parents in-law Mr. Venkatanarayana and Mrs. Varalakshmi for their constant support and encouragement throughout the tenure of research work.

I avail this opportunity to thank my parents who are the Almighty’s most treasured gift to me. I feel scanty of words to magnitude the boundless love and tireless sacrifice and affection showed on me by my parents, Mr. Nagesh and Mrs. Nagaveni and affection of my sister Savita and brother Manjunath that I could attain academic heights to accomplish my doctoral degree. And I express my deepest adoration to them for teaching me the etiquettes of life.

The Council of Scientific and Industrial Research (CSIR), New Delhi, India is greatly acknowledged for providing me fellowship to carry out my research work. I also acknowledge the financial support from Generation Challenge Programme (GCP) to the ADOC and Tropical Legumes1 (TLI) projects where I worked in.

I convey my whole hearted thanks to many of my well wishers and other friends requesting their forgiveness for not mentioning them here by name. Date: Place: Hyderabad (Spurthi N Nayak)

CONTENTS

Chapter No Title Page No

1 Introduction 1

2 Review of literature 5

3 Materials and methods 48

4 Results 69

5 Discussion 151

6 Summary 171

7 Literature 175

Appendices 209

ABBREVIATIONS

˚C : degree Celsius

µl : microliter

ABA : Abscisic acid

ADOC : Allelic Diversity at Orthologous Candidate genes

AFLP : Amplified Fragment Length Polymorphism

ANOVA : Analysis of Variance

ASR : Abscisic acid Stress and Ripening

BAC : Bacterial Artificial Chromosome

BES : BAC-end Sequences

BIBAC : Binary Bacterial Artificial Chromosome

BLAST : Basic Local Alignment Search Tool

bp : base pair

CaM : Cicer arietinum Microsatellites

CAP2 : Cicer Apetala gene 2

CAPS : Cleaved Amplified Polymorphic Sequences

cDNA : complementary DNA

CIM : Composite Interval Mapping

cM : centiMorgan

COS : Conserved Orthologous Set

CTAB : Cetyl Trimethyl Ammonium Bromide

DArT : Diversity Array Technology

dCAPS : derived Cleaved Amplified Polymorphic Sequences

DDBJ : DNA Data Bank of Japan

DIVEST : Diversity Estimator

DNA : Deoxyribonucleic Acid

DREB : Drought Responsive Element Binding proteins

EMBL : European Molecular Biology Laboratory

EREBP : Ethylene Responsive Element Binding Protein

EST : Expressed Sequence Tag

GCP : Generation Challenge Programme

GSS : Genome Survey Sequences

ICCM : ICRISAT Chickpea Microsatellites

ICRISAT : International Crops research Institute for the Semi-Arid Tropics

Indel : Insertion deletion

INRA-CNG : Institut National de la Recherche Agronomique-Centre National de

Génotypage

ISSR : Inter Simple Sequence Repeats

JCVI : J. Craig Venter Institute

kbp : kilo base pairs

LD : Linkage Disequilibrium

LDW : Leaf Dry Weight

LG : Linkage Group

LOD : Logarithm of odds (base 10)

LnPD : Natural logarithm of probability of data

MAS : Marker-Assisted Selection

Mb : Million bases

MISA : MIcroSAtellite

mM : milliMolar

MSA : Multiple Sequence Alignment

NCBI : National Center for Biotechnology Information

NIPGR : National Institute for Plant Genome Research

ng : nanograms

NLS : Nuclear Localization Sequence

PAGE : Polyacrylamide Gel Electrophoresis

PCR : Polymerase Chain Reaction

PIC : Polymorphism Information Content

QTL : Quantitative Trait Loci

RAPD : Random Amplified Polymorphic DNA

RD : Rooting Depth

RDW : Root Dry Weight

RFLP : Restricted Fragment Length Polymorphism

RGA : Resistance Gene Analogues

RIL : Recombinant Inbred Line

RL : Root Length

RLD : Root Length Density

RSA : Root Surface Area

RT : Root Dry Weight/Total Dry Weight ratio

RV : Root Volume

SCAR : Sequence Characterized Amplified Region

SDW : Shoot Dry Weight

SIM : Simple Interval Mapping

SNP : Single Nucleotide Polymorphism

SMA : Single Marker Analysis

SPS : Sucrose Phosphate Synthase

SSCP : Single Strand Conformational Polymorphism

SSR : Simple Sequence Repeats

SSRIT : Simple Sequence Repeat Identification Tool

StDW : Stem Dry Weight

STMS : Sequence Tagged Microsatellite Sites

SuSy : Sucrose Synthase

TIGR : The Institute for Genome Research

LIST OF TABLES

Table No.

Title Page No.

1 Microsatellite markers available in chickpea 18 2 A list of some genetic maps in chickpea 23 3 Some QTL studies related to drought tolerance in selected crop

species 29

4 QTL studies related to agronomic traits in chickpea 30 5 List of some candidate genes involved in drought tolerance 37 6 Primers used to amplify candidate genes related to drought tolerance 65 7 Novel set of SSR markers (ICCM) developed from SSR-enriched

library of chickpea 71

8 Functional annotation of ICCM sequences with EST databases 80 9 BLASTN results of forty ICCM sequences showing significant

similarity with Hologalegina across nine plant databases 81

10 Novel set of SSR markers (CaM) derived from BAC-end sequences in chickpea

85

11 Distribution of different types of mapped marker loci on different linkage groups (LGs) of the inter-specific map of chickpea

120

12 Distribution of mapped marker loci on different linkage groups of the inter-specific map of chickpea

120

13 Distribution of mapped SSR marker loci on different linkage groups of the intra-specific map of chickpea

122

14 Analysis of variance (ANOVA) for drought tolerance related root traits in chickpea

123

15 Two-sided test of correlation studies across ten drought tolerance related root traits using GenStat

125

16 Heritability of the drought tolerance related root traits 125 17 List of major QTLs explaining >20% phenotypic variation 127 18 Main effect QTLs for drought tolerance related root traits using

single locus analysis using QTL cartographer 128

19 Descriptive statistics for drought tolerance related root traits in the chickpea reference set

139

20 Correlation coefficients among drought tolerance related root traits in chickpea

139

21 Summary of sequence diversity of candidate genes across eight diverse chickpea genotypes

141

22 Estimates on sequence diversity in the chickpea reference set for five candidate genes

144

23 Association of candidate genes with drought tolerance related traits 147 24 Association of DArT markers with drought tolerance related traits

based on genome-wide association approach 149

LIST OF FIGURES

Figure No.

Title In between page nos.

1 Frequency of different SSR classes in sequences obtained from SSR enriched library

78-79

2 Distribution of microsatellites with varying repeat units in SSR enrichment library

78-79

3 Distribution of Class I and Class II repeats in newly isolated ICCM microsatellites

79-80

4 Phylogenetic relationships of Papilionoideae family 82-83 5 Frequency of different SSR classes in BAC end sequences 83-84 6 Distribution of microsatellites with varying repeat units in BAC-end

sequences 83-84

7 Distribution of Class I and Class II repeats in newly isolated CaM microsatellites

84-85

8 Inter-specific genetic map of chickpea derived based on mapping populations- C. arietinum ICC 4958 × C. reticulatum PI 489777

118-119

9 Intra-specific map of chickpea derived based on mapping population- C. arietinum ICC 4958 × C. arietinum ICC 1882

121-122

10 QTL map for drought tolerance related traits based on intra-specific mapping population- ICC 4958 × ICC 1882

127-128

11 A snapshot of “QTL hot spot region” located on LG6 of intra-specific mapping population- ICC 4958 × ICC 1882

137-138

12 Multiple sequence alignment (MSA) of CAP2 promoter across eight chickpea genotypes along CAP2 promoter sequence

141-142

13 MSA of abscisic acid stress and ripening (ASR) gene across eight chickpea genotypes along with one Medicago genotype (A17)

142-143

14 MSA of sucrose synthase (SuSy) gene across eight chickpea genotypes along with one Medicago genotype (A17)

142-143

15 MSA of sucrose phosphate synthase (SPS) gene across eight chickpea genotypes along with one Medicago genotype (A17)

142-143

16 Haplotype network of ASR gene developed based on country of origin of genotypes of chickpea reference set

146-147

17 Haplotype network of SPS gene developed based on country of origin of genotypes of chickpea reference set

146-147

18 Haplotype network of fragments of ERECTA gene developed based on country of origin of genotypes of chickpea reference set

146-147

19 Estimation of delta K value using Evanno’s method 147-148 20 Structure plot of reference set of chickpea at K=11 147-148

ABSTRACT

Low levels of polymorphism and lack of sufficient numbers of molecular markers such as

microsatellite or simple sequence repeats (SSRs) are the main constraints in chickpea

improvement. Hence to increase the number of SSR markers, 1,655 novel SSRs were

developed from SSR-enriched genomic library (311) and mining the BAC-end sequences

(1,344). These markers, along with already available markers were tested for

polymorphism on parental genotypes of the inter-specific (ICC 4958 × PI 489777) and

intra-specific mapping population (ICC 4958 × ICC 1882). As a result, a comprehensive

inter-specific genetic map of 621 marker loci, spanning a genetic distance of 984.11cM

was prepared. For identification of QTLs for drought tolerance traits, an intra-specific

map (segregating for drought tolerance related traits) consisting of 230 SSR loci,

spanning 466.95cM genetic distance was constructed after screening 2,409 SSR markers.

The QTL analysis detected 47 significant QTLs for the ten root traits, of which seven

were major QTLs (>20% phenotypic variation). The QTL analysis revealed the presence

of a “QTL hot-spot” region explaining 49.9% phenotypic variation was detected. For

undertaking association mapping for drought tolerance, two approaches namely candidate

gene sequencing and genome-wide scanning approaches were used on the reference set

comprising of 318 chickpea genotypes. In case of the candidate gene sequencing

approach, five candidate genes associated with drought tolerance were selected namely,

chickpea Apetala2 (CAP2-which is the homolog of DREB2A), abscisic acid stress and

ripening hormone (ASR), sucrose synthase (SuSy), sucrose phosphate synthase (SPS) and

ERECTA genes. Highest nucleotide diversity was observed in case of ERECTA followed

by ASR gene and the lowest for CAP2 gene. Association analysis based on candidate gene

sequencing showed the association of two genes (ASR and CAP2 promoter) with drought

tolerance related traits. Apart from this, the genome-wide association studies using 1,157

DArT markers showed the significant association of 26 DArT markers with eight drought

tolerance related traits.

In summary, developed genomic resources such as SSR markers and genetic maps will be

useful for chickpea genetics and breeding applications. Moreover, markers and genes

associated with QTLs for drought tolerance related traits will be useful for molecular

breeding for drought tolerance in chickpea improvement.

1

1. INTRODUCTION

Chickpea (Cicer arietinum L.) is a self-pollinated diploid (2n = 2x = 16) annual legume with

a genome size of ~738 Mbp (Arumuganathan and Earle 1991). Chickpea (Cicer arietinum

L.), the third most important cultivated grain legume in the world after soybean and beans

(FAOSTAT 2009). It is the member of the Leguminosae family, which includes 18,000

species, grouped into 650 genera (http://www.ildis.org/Leguminosae). Chickpea is

commonly called as gram, Bengal gram or garbanzo bean is mainly cultivated in the Indian

subcontinent, West Asia, the Mediterranean, North Africa and the America (Croser et al.

2003). Over 95% of chickpea production area and consumption occur in developing

countries, most of which are rainfed, with India contributing the largest share (71%),

followed by Pakistan (10%), Iran (5%) and Turkey (4%) (FAOSTAT database-

http://faostat.fao.org/site/567/default.aspx#ancor, 2009).

Among grain legumes, chickpea is rich in nutritional compositions and does not contain

significant quantities of any specific major anti-nutritional factors. On an average, chickpea

seed contains 23% of highly digestible protein, 64% crude fiber, 6% soluble sugar and 3%

ash. The mineral component is high in phosphorous (343mg/100g), calcium (186mg/100g),

magnesium (141mg/100g), iron (7mg/100g) and zinc (3mg/100g) (Williams and Singh

1987). There are two market types of cultivated chickpea namely, desi (with small and

brown seeds) and kabuli (with bold and cream coloured seeds). In India, desi type of

chickpea, accounts for nearly 90% of total area under cultivation and remaining (10%)

cultivated area being occupied by kabuli type (Pundir et al. 1985).

Chickpea is a hardy, deep-rooted dry land crop and can grow to full maturity despite

conditions that would prove fatal for most of the crop plants. It is grown on marginal land

and rarely receives fertilizers or protection from diseases and insect pests (Singh and Reddy

2

1991). Nearly 90% of the crop is cultivated under rainfed conditions mostly on receding soil

moisture, the region is called semi-arid tropics (SAT). Currently, production of chickpea is

very low (6.5 mt, FAO 2009) and has stagnated for past several years. This can be attributed

to a number of biotic and abiotic stresses that reduced the yield and yield stability, mainly

Ascochyta blight, Fusarium wilt, Helicoverpa pod borer, Botrytis grey mold, drought and

cold. The estimated yield losses due to abiotic stress (6.4 mt) are much more than loss due

to biotic stress (4.8 mt) (Ryan 1997). Among various factors effecting yield in chickpea,

drought is a major constraint causing 40-50% reduction in chickpea yield globally (Ahmad

et al. 2005). Considering the constraining issues and their relative affect on the global yield,

it is very important and timely to work for the improvement of drought and salinity

tolerance for stabilization of the yield in chickpea. Therefore, improving resistance to biotic

and tolerance to abiotic stresses as well as a general increase in dry matter are major aims of

chickpea breeders around the world.

In both Mediterranean and sub-tropical climates, seed filling in chickpea is subject to

terminal drought, which limits seed yield (Turner et al. 2001). Many physiological processes

associated with crop growth and development is reported to be influenced by water deficits

(Turner and Begg 1978). Breeding for drought tolerance is generally considered slow due to

the quantitative and temporal variability of available moisture across years, the low

genotypic variance in yield under these conditions, inherent methodological difficulties in

evaluating component traits (Ludlow and Muchow 1990), together with the highly complex

genetic basis of this character (Turner et al. 2001). Genetic improvement of drought

tolerance, therefore, is a challenge using conventional breeding approaches that rely on

selection for yield in drought-stressed environments. However, the large genotype by

environment (G × E) interactions for yield and the difficulties of controlling the level of

water stress under natural conditions makes direct selection for yield ineffective. Thus, the

3

application of a holistic approach, combining genomics with breeding and physiology,

termed as genomics-assisted breeding (Varshney et al. 2005b), provides strategies for

improving component traits of drought tolerance that should prove more effective and

efficient than the conventional selection methods. It is essential to identify quantitative trait

loci (QTLs) or genes that confer drought tolerance so that these QTLs/genes can be

deployed in breeding programmes. There are two main approaches in detecting the QTLs

associated with the trait of interest- one being done by “linkage mapping” or “QTL interval

mapping” and the other is based on linkage disequillibrium (LD) called as “linkage

disequilibrium mapping” or “association mapping”.

Due to relatively low levels of polymorphism between cultivated chickpea genotypes, inter-

specific crosses between C. arietinum and C. reticulatum have been the primary focus for

genetic studies of agronomic traits (Singh et al. 2008). A diverse array of technologies is

available to identify and monitor DNA polymorphism and as a consequence molecular

markers are now routinely used in the breeding programs of several crop species (Varshney

et al. 2006, 2007). Various kinds of molecular markers such as restriction fragment length

polymorphism (RFLP), random amplified polymorphic DNA (RAPD), amplified fragment

length polymorphism (AFLP), microsatellites or simple sequence repeat (SSR), sequence

tagged site (STS) and single nucleotide polymorphism (SNP) markers have become

available for characterization of plant genome structure and genetic improvement in several

crops species through construction of genetic maps, gene tagging and marker-assisted

selection (MAS) (Mohan et al. 1997; Gupta and Varshney 2000; Varshney et al. 2005b;

Rafalski et al. 2002). Until recently, SSR markers have been proven as the markers of

choice for plant breeding applications because of their wide genomic distribution, technical

simplicity, amenability to high-throughput automation, co-dominant inheritance, high

reproducibility, multi-allelic nature and chromosome-specific location. Very recently, SNP

4

markers have also started to emerge as informative and useful markers for chickpea genetics

and breeding.

In case of chickpea, unlike other plant species, the progress in development of molecular

markers has been very slow. Development of SSR markers was initiated only in late 90s. As

a result, several hundred SSR markers were available at the time of undertaking this study

(Choudhary et al. 2006; Hüttel et al. 1999; Lichtenzveig et al. 2005; Sethy et al. 2003,

2006a, 2006b; Winter et al. 1999). However, low level of polymorphism in chickpea

germplasm has limited integration of SSR markers into chickpea genetic maps. Therefore, it

is ascertained that there is need to increase the density of markers on the genetic maps and

study marker-trait association using different approaches.

Keeping the above in view, the present study is proposed with the following objectives:

1) Development of novel set of microsatellite or simple sequence repeat (SSR) markers

in chickpea

2) Integration of novel SSR markers into the reference genetic map using a mapping

population derived from the cross C. arietinum (ICC 4958) × C. reticulatum

(PI 489777)

3) Construction of an intra-specific (C. arietinum) genetic map using a mapping

population derived from the cross ICC 4958 × ICC 1882 and identification of QTLs

for drought tolerance related traits by using linkage mapping approach

4) Identification of candidate genes for drought tolerance through comparative

genomics and bioinformatics approaches and study of allelic sequence diversity

(single nucleotide polymorphisms) in a reference set of chickpea

5) Identification of genes/markers associated with drought tolerance using candidate

gene sequencing and genome-wide association approaches

5

2. REVIEW OF LITERATURE

2.1 Chickpea

The genus Cicer (family Fabaceae) consists of 43 species of which 9 are annual including

cultivated chickpea, 33 are perennial and one is unclassified. The annual species have been

classified into two sections, Monocicer and Chamaecicer, based on morphological

characters, life cycle and geographical distribution (van der Maesen 1987). Eight of the

annual species (C. arietinum, C. reticulatum, C. echinospermum, C. bijugum, C. judaicum,

C. pinnatifidum, C. yamashitae and C. cuneatum) belong to the section Monocicer, while

section Chamaecicer contains the annual species, C. chorassanicum.

Chickpea (C. arietinum L.) is one of the pulse crops domesticated in the Old World about

7000 years ago. Cytogenetic and seed protein analyses are consistent with C. reticulatum as

the wild progenitor of domesticated C. arietinum, with south-eastern Turkey as the

presumed centre of origin (Ladizinsky and Adler 1976). This claim was supported by van

der Maesen (1987) based on the presence of the closely related annual species, C.

reticulatum and C. echinospermum in southeastern Turkey. Three wild annual Cicer

species, C. bijugum, C. echinospermum and C. reticulatum, closely related to cultivated

chickpea, cohabit in this area. The cultivated species, C. arietinum is found only in

cultivation and cannot colonize successfully without human intervention. On the other hand,

the wild species (e.g. C. reticulatum, C. bijugum) occur in weedy habitats (fallow or

disturbed habitats, roadsides, cultivated fields of wheat, and other places not touched by

man or cattle), mountain slopes among rubble (e.g. C. pungens, C. yamashitae), and on

forest soils, in broad-leaf or pine forests (e.g. C. montbretii, C. floribundum).

6

Cultivated chickpea (C. arietinum) is composed of two genetically distinct sub-types that

are readily distinguished based on seed size and colour: desi, composed of small, angular,

brown seeded varieties, and kabuli, composed of large, smooth, cream seeded varieties.

Regarding the origin of kabuli and desi types of chickpea, it is reported that desi types

originated first, followed by kabuli type which was developed by selection and mutation

(Singh 1997). There is linguistic evidence that kabuli type reached India via Kabul, the

capital of Afghan, about two centuries ago and acquired the name as ‘kabuli chana’ in Hindi

(van der Maesan 1972). In India, desi type of chickpea accounts for nearly 90% of total area

under chickpea cultivation and remaining area being occupied by kabuli type. Across the

world, the ratio of desi and kabuli chickpea production is 3:1. In addition `gulabi', pea

shaped forms of local importance are also recognized. Kabuli types are considered relatively

more advanced because of their larger seed size and reduced pigmentation achieved through

conscious selection. Study at International Crops Research Institute for the Semi-Arid

Tropics (ICRISAT) revealed that desi and kabuli types differ in their dietary fiber

components of seed both qualitatively and quantitatively (Singh 1985). For instance, kabuli

types, as compared to desi types contain higher amount of dietary fiber, particularly

cellulose and hemicellulose.

2.2 Constraints in Chickpea Production

Chickpea is a hardy, deep-rooted dryland crop and can grow to full maturity despite

conditions that would prove fatal for most crops. It is grown on marginal land and rarely

receives fertilizers or protection from diseases and insect pests (Singh and Reddy 1991).

Despite of its economical importance chickpea productivity is low because of yield loses

due to abiotic (drought, cold and salinity) and biotic stresses (Helicoverpa, Ascochyta blight,

Fusarium wilt and Botrytis grey mold). Chickpea is grown mostly as a post-rainy season

crop on soil moisture conserved from the preceding rainy season. The crop is therefore,

7

often exposed to terminal drought and heat stress. The estimated yield losses due to abiotic

stress (6.4 mt) are much more than loss due to biotic stress (4.8 mt) (Ryan 1997). Among

various factors effecting yield in chickpea, drought is a major constraint causing 40-50%

reduction in chickpea yield globally (see Ahmad et al. 2005). With climate change and

predictions of increased water scarcity in the future, water-stress is likely to remain as a

primary constraint. As irrigation is often not available in the semi-arid tropics (SAT),

especially for resource poor farmers, it is critical that genetic enhancement focus on

maximizing the extraction of available soil moisture and improving the efficiency of water

use in crop establishment, growth and yield.

2.3 Drought Stress in Chickpea

Drought, cold and salinity are the major abiotic stresses affecting chickpea in order of their

importance (Croser et al. 2003). As drought is the major constraint in chickpea production,

it is essential to develop varieties which can make use of available water resources and

produce maximum yield. Drought is defined as a water stress due to lack or insufficient

rainfall and/or inadequate water supply (Toker et al. 2007). The seriousness of drought

stress depends on its timing, duration and intensity (Serraj et al. 2003). Often drought is

accompanied by relatively high temperatures, which promote evapotranspiration and hence

could accentuate the effects of drought and thereby further reduce crop yields.

Effect of drought stress depends on evapotranspiration, soil water holding capacity and the

crop water requirements (Toker et al. 2007). Worldwide, 90% of chickpea is grown under

rainfed conditions (Kumar and Abbo 2001) where terminal drought is one of the major

constraints limiting crop productivity (Toker et al. 2007). Terminal drought is a usual

feature in semi-arid tropics like south Asia and northern Australia where chickpea is grown

in the post-rainy season on progressively receding soil moisture conditions (Leport et al.

8

1998; Siddique et al. 2000). The damage due to drought is compounded by heat stress in the

warmer Mediterranean regions and regions like South Asia where temperature increases

towards flowering (Singh et al. 1997) and it is difficult to differentiate between the damage

caused by the individual stresses. As a result of drought stress, the growing season may be

shortened affecting yield components, i.e., total biomass, pod number, seed number, seed

weight and quality, and yield per plant (Toker et al. 2007). Flowering and seed set are the

most critical growth stages affected by drought in chickpea (Khanna-Chopra and Sinha

1987).

2.3.1 Breeding for drought tolerance

Conventional breeding for drought tolerance is based on the selection for yield and its

components under a water-limited environment (Blum 1985; Millan et al. 2006). The

germplasm is usually screened for two important drought avoidance/tolerance traits; large

root system and small leaf area (Turner et al. 2001; Saxena 2003). Previously, more than

1500 chickpea lines were screened for drought tolerance and the genotype ICC 4958 was

the most promising for drought tolerance (Saxena et al. 1993). Subsequently, ICC 4958 was

used in a three-way cross with cv. Annigeri and the Fusarium wilt resistant genotype ICC

12237. The progenies were selected for high yield and drought tolerance traits (Saxena

2003). Several lines combining the large root traits of ICC 4958 and the small leaf area traits

of ICC 5680 were reported to be more drought tolerant and yielded similarly to the high

yielding parent (Saxena 2003). In another study, a chickpea mini-core germplasm collection

comprising 211 genotypes along with 12 popular cultivars and 10 annual wild chickpea

genotypes was screened for root traits. Interestingly, several C. arietinum genotypes with

more root depth than ICC 4958 were also identified, which could serve as an alternative

source for the large root trait. An outstanding genotype, ICC 8261, which had the largest

RLD and deepest root system, was identified in chickpea mini-core germplasm collection

9

(Kashiwagi et al. 2005; Krishnamurthy et al. 2003). Also, genotypic variation for osmotic

adjustment in chickpea was studied across 120 mixed F2 population obtained from crosses

Kaniva × Tyson and Kaniva × CTS60543. The correlation of osmotic adjustment with yield

under drought stress is unclear and the heritability was low (h2 = 0.20- 0.33) (Morgan et al.

1991; Turner et al. 2001; Moinuddin and Khanna-Chopra 2004). In summary, breeding for

drought tolerance has remained hampered due to limited knowledge about the genetic basis

of drought tolerance and its negative correlations with productivity (Mitra 2001). Moreover,

selection for yields in chickpea is not effective in early segregating generations because of

its indeterminate growth habit. Therefore breeders have to select for crosses rather than

plants in F2 and F3 generations (Ahmad et al. 2005).

2.3.2 Molecular breeding for drought tolerance

Recent years have witnessed tremendous progress in the development of novel genetic tools

and genomics approaches including development of molecular markers, dense genetic maps

and whole-genome transcription profiling techniques to identify genomic regions and genes/

quantitative trait loci (QTLs) underlying plant stress responses in many crop species (see

Varshney et al. 2005a). The development of molecular techniques for genetic analysis has

led to a great increase in our knowledge of crop genetics and genomics. Improvement in

marker detection systems and the techniques used to identify markers-trait associations

based on complete linkage maps and bulked segregant analysis has made great advances in

recent years. Plant breeding has benefited from DNA marker technologies that were used to

establish saturated genetic maps in major crop species including cereals (Varshney et al.

2006) and legumes (Varshney et al. 2010). Markers in a high density genetic map will allow

the precise tagging of mono- and oligo-genic traits, with the dual goal of marker-assisted

selection for traits and positional cloning of the underlying genes (Tanksley et al. 1995).

Use of genomic tools like molecular markers and other tools in integrated approach for crop

10

improvement has also been referred as “genomics- assisted breeding” (Varshney et al.

2005b).

Mapping of the chickpea genome has been of interest to identify genomic locations of

disease resistance genes and other yield related traits (Winter et al. 2000; Cho et al. 2002;

Rajesh et al. 2002a, 2004; Abbo et al. 2005). However, due to very low polymorphisms in

cultivated chickpea gene pool (Udupa et al. 1993; Labdi et al. 1996), progress in chickpea

genomic research has been relatively slow compared with other legume species (see

Varshney et al. 2010). Moreover for drought tolerance related traits, very limited efforts

have been made in past in the area of molecular mapping. As mentioned above, root traits

(e.g. root biomass, root volume, etc.) have been considered as target traits for breeding for

drought tolerance. Since selection of root traits is very laborious, molecular tagging of major

genes/ QTL for these traits may enable marker-assisted selection (MAS) and greatly

improve the precision and efficiency of breeding (see Millan et al. 2006; Varshney et al.

2010). In this context, one RIL population from Annigeri (less root biomass) × ICC 4958

(high root biomass) was screened to identify molecular markers for root traits. Because of

availability of only a few hundred SSR markers in public domain at the time of undertaking

this study, genotyping data could be generated for only 14 polymorphic SSR markers.

Therefore, a good genetic map couldn’t be developed and single marker linear regression

analysis showed association of SSR marker TAA170 with relatively high phenotypic

variation for drought tolerance related traits like root length (R2 = 33.1%), root weight (R2 =

33.1%) and shoot weight (R2 = 54.2%), where R2 was the adjusted coefficient of

determination (Chandra et al. 2004). With an objective to develop diverse mapping

populations, contrasting parental genotypes were selected after screening the mini core

collection for root traits (Kashiwagi et al. 2005). As a result, two RIL populations ICC 4958

×ICC 1882 and ICC 283 × ICC 8261 have been developed and these showed good

segregation for root traits.

11

2.4 Chickpea Genomics

2.4.1 Molecular markers

Molecular markers can be defined as the DNA segements that show differences in the

nucleotide sequences of DNA of one or more individuals at corresponding sites on the

homologous chromosomes that follow a simple Mendelian pattern of inheritance. Molecular

markers offer numerous advantages over conventional morphological and biochemical

markers as they are stable, virtually unlimited in number, detectable at all plant tissues

regardless of growth, differentiation and development, not confounded by environment and

having no pleiotropic and epistatic effects (Bennett 1994). These are the powerful tools for

the fingerprinting, genetic diversity analysis, tagging of useful genes, construction of

genetic and physical maps, identification of QTLs, marker-assisted selection (MAS), gene

pyramiding, evolutionary studies and positional cloning of useful genes (Koebner et al.

2001; Gupta and Rustgi 2004; Varshney et al. 2005b; Agarwal et al. 2008). A number of

molecular marker technologies have been utilized to visualize DNA polymorphism in

several crop species (Staub et al. 1996; Gupta et al. 2002; Jones et al. 2009).

Depending on the method of detection of the sequence variation, the molecular markers

have been classified as hybridization based (PCR-independent) molecular markers, PCR-

dependent molecular markers and micro-array based molecular markers (Gupta et al. 2002).

Hybridization based molecular markers include restriction fragment length polymorphism

(RFLP). PCR-dependent molecular markers include random amplified polymorphic DNA

(RAPD), amplified fragment length polymorphism (AFLP), simple sequence repeats (SSR)

or microsatellite, sequence tagged sites (STS) and cleaved amplified polymorphic sequence

(CAPS) (Gupta et al. 2002, Semagn et al. 2006). Micro-array based molecular markers

comprise of single nucleotide polymorphism (SNP), single feature polymorphism (SFP) and

diversity array technology (DArT) (Gupta et al. 2008).

12

2.4.1.1 Hybridization based molecular markers

Hybridization based molecular markers were the first molecular markers used in genomics

research in animal or plant systems. Restriction fragment length polymorphism (RFLP) was

the first marker system of this category and was based on differences in the recognition sites

for restriction enzymes in the genome of the species.

RFLP refers to difference between two or more samples of homologous DNA molecules

arising from differing locations of restriction sites, and to a related laboratory technique by

which these segments can be distinguished. In RFLP analysis, the DNA sample is digested

by restriction enzymes and the resulting restriction fragments are separated according to

their lengths by gel electrophoresis. Although now largely obsolete, RFLP analysis was the

first DNA profiling technique inexpensive enough to see widespread application. In addition

to genetic fingerprinting, RFLP was an important tool in genome mapping, localization of

genes for genetic disorders, determination of risk for disease, and paternity testing. RFLP is

the first molecular marker to be developed and used for human (Botstein et al. 1980) and

plant (Weber and Helentjaris 1989) genome mapping. RFLP markers are highly

reproducible, co-dominant, locus specific markers, which make them ideal marker system

for plant genome analysis. In fact, in chickpea also, the first genetic map constructed in

chickpea constituted RFLP and RAPD markers along with isozyme markers (Simon and

Muehlbauer 1997). Genetic diversity studies were also carried out using RFLP markers

(Udupa et al. 1993) and microsatellite-derived RFLP markers (Sharma et al. 1995; Serret et

al. 1997) that though showed the narrow genetic variability in chickpea cultivars.

2.4.1.2 PCR-based molecular markers

The invention of polymerase chain reaction (PCR) by K. Mullis in year 1983 revolutionized

the molecular genetics research including molecular markers. As a result, a large number of

13

approaches for generation of molecular markers based on PCR have become available,

primarily due to its apparent simplicity and high probability of success in giving

amplification of large number of copies of target DNA. PCR-based marker systems have

been further divided into two categories, (1) arbitrarily primed (non-sequence specific)

which include RAPD, AFLP marker systems and (2) sequence tagged PCR based technique

include STS, CAPS and SSRs.

2.4.1.2.1 Random amplified polymorphic DNA (RAPD)

RAPD is a dominant marker based PCR technique, which employs single decamer primer of

arbitrary sequence sequences for differential amplification of genomic DNA from the

genome (Williams et al. 1990). The polymorphism in the RAPD profile is produced by

rearrangement or deletion at or between oligonucleotide primer binding sites in the genome,

which causes absence or presence of a band in gel electrophoresis (Rafalski and Tingey

1993). This method is quick, simple and requires less amount of DNA per assay. Being

based on random primers, it does not require prior knowledge of sequence information for

its design. A large number of such primers are commercially available that work across

organisms starting from bacteria to human. Due to these advantages, RAPD marker system

has been used more frequently for genetic diversity studies, variety identification and

understanding genetic relationships in crop species.

In case of chickpea also, several studies have been conducted by using RAPD markers. For

instance, Ahmad (1999) and Sudupak et al. (2002) used RAPD markers to investigate

genetic relationships among the annual Cicer species. The RAPD analysis placed C.

arietinum, C. reticulatum and C. echinospermum in a single cluster, C. yamashitae and C.

chorassanicum in the next cluster, C. pinnatifidum, C. bijugum and C. judaicum in third

cluster and C. cuneatum in the fourth cluster.

14

Use of dominant RAPD markers can be enhanced with identification of coupling and

repulsion phase markers linked to the gene of interest. The coupling and repulsion markers

can be used together as a co-dominant pair and will be equally informative as co-dominant

markers in detecting heterozygotes in segregating populations (Haley et al. 1994; Johnson et

al. 1995; Singh et al. 2008). A major limitation of this marker system is dominant

inheritance and non-reproducibility due to low annealing temperature. However, utility of

desired RAPD marker can be increased by converting it into more reproducible informative

marker (Paran and Michelmore 1993) termed as sequence characterized amplified region

(SCAR).

2.4.1.2.2 Amplified fragment length polymorphism (AFLP)

In order to overcome the limitations of reproducibility associated with RAPD, AFLP marker

system (Vos et al. 1995) was developed by selective amplification of DNA fragments

obtained by restriction enzyme digestion. It combines the power of RFLP with the

flexibility of PCR-based technology by ligating primer-recognition sequences (adaptors) to

the restricted DNA and selective PCR amplification of these restriction fragments using a

limited set of primers labelled either with radioisotope and/ or fluorescent dye. The

amplified products thus can be resolved on sequencing gels/capillaries and visualized by

autoradiography or by laser based scanning in an automated fragment analysis system.

About 50 to 100 amplified fragments are obtained per assay in AFLP among all the DNA

profiling systems, which increases its probability of detecting polymorphism many folds

(Thomas et al. 1995; Joseph et al. 2004).

AFLP marker system has been used for genetic diversity estimation in cultivated chickpea

and its wild relatives in order to discover the origin and history of chickpea (Nguyen et al.

2004; Talebi et al. 2008, 2009). The requirement of significant technical skills, laboratory

facilities, financial resources and high quality genomic DNA for complete restriction

15

digestion and dominant inheritance has limited the use of AFLP markers in plant genetic

analysis.

2.4.1.2.3 Cleaved amplified polymorphic sequence (CAPS)

The CAPS marker technique provides a way to utilize the DNA sequence of mapped RFLP

markers by eliminating the tedious DNA blotting and thus is known as PCR-AFLP markers

(Konieczny and Ausubel 1993). The CAPS assay is performed by digesting locus-specific

PCR amplicons with one or more restriction enzymes followed by separation of digested

DNA fragments on agarose or polyacrylamide gels. It deciphers the restriction fragment

length polymorphism resulting from the single base substitution like SNPs and

insertions/deletions, which modify the recognition sites of the restriction enzymes

(Chelkowski and Stepien 2001). The primers are designed based on prior sequence

information of genomic and cDNA sequences and cloned RAPD amplicons. The CAPS

analysis is versatile and can be combined with single strand conformational polymorphism

(SSCP), SCAR, AFLP and RAPD analysis to increase the possibility of finding DNA

polymorphism. It is robust and cost-effective assay that can be implemented in laboratories

lacking sufficient infrastructural facilities. CAPS markers are characterized by their co-

dominant inheritance and locus specific nature and useful for genotyping applications

(Parsons and Heflich 1997; Weiland and Yu 2003). However this technique works only

when SNP site is present at the recognition site of restriction enzyme, while those changes

outside the target site of enzyme cannot be assayed by CAPS analysis. In order to overcome

this, a variant of the CAPS method called dCAPS (derived-CAPS) has been developed by

Neff et al. (2002). In dCAPS analysis, the conversion of mutation sites into CAPS markers

by the artificial introduction of restriction sites involves the creation of mismatched PCR

primers. Its successful application is not always trivial depending on the number, positions

and types of mismatches. A computer-based software tool ‘SNP2CAPS’ has also been

16

developed (Thiel et al. 2004) in order to deduce CAPS candidates from the sequences

having SNPs.

In case of chickpea, CAPS and dCAPS markers have been developed from bacterial

artificial chromosome (BAC)-end sequences (Rajesh et al. 2005) and EST sequences

(Varshney et al. 2007c). The CAPS and dCAPS-RGA markers have also been integrated

into composite genetic map of chickpea and their association with disease resistance was

studied (Palomino et al. 2009).

2.4.1.2.5 Simple sequence repeats (SSRs) or Microsatellites

Microsatellite markers are also known as simple sequence repeats or SSRs or sequence-

tagged microsatellite site (STMS) (Beckmann and Soller 1990) consisting tandem repeats of

1-6 bp in length (Gupta and Varshney 2000). They consist of head-to-tail tandem arrays of

short DNA motifs and are often highly polymorphic due to variable numbers of repeat units.

Microsatellites as DNA markers are advantageous over many other markers as they are

highly polymorphic, highly abundant, analytically simple and readily transferable (Weber

1990). SSR loci show high levels of length polymorphism, on account of differences in the

number of repeat units, and show co-dominance. Microsatellites are reported to be more

variable than RFLPs and RAPDs, and have been widely utilized in plant genomic studies

(Gupta and Varshney 2000). The advantages of microsatellite markers over other types of

genetic markers will become more important, and more obvious, when they are used to

track desirable traits in large-scale breeding programs and as anchor points for map-based

gene cloning strategies (Brown et al. 1996). Because of its cost and time effectiveness, use

of microsatellite markers in plant breeding has become a routine procedure in several crop

species mainly cereals (Varshney et al. 2006) and also some legumes (Varshney et al. 2010).

In late 90s, microsatellites or simple sequence repeats (SSR) have come to use in plant

17

genetics, and have proved as the molecular marker system of choice for many areas of

genome analyses and breeding applications (Dib et al. 1996; Powell et al. 1996).

Gel hybridization of restriction-digested genomic DNA with microsatellite specific probes

revealed polymorphic fingerprints in chickpea (Weising et al. 1992; Sharma et al. 1995a).

However, the applicability of this technique for mapping was limited by a high incidence of

non-parental bands in progenies, and by the finite number of polymorphisms detected

between cultivars (~80 bands; Sharma et al. 1995a). Multi-locus banding patterns were also

obtained when microsatellite-complementary oligonucleotides were used as PCR primers to

amplify the chickpea genome, but intraspecific variation was again low (Sharma et al.

1995a; Ratnaparkhe et al. 1998). Another technique based on microsatellites employed

sequence information of repeat-flanking regions to design locus-specific PCR primer pairs

for amplifying the SSR locus and detecting the polymorphism (Weber 1990; Akkaya et al.

1992; Morgante and Olivieri 1993). This technique often referred as STMS/ SSR

polymorphism. This technique has been used for polymorphism detection in several crop

species (see Powell et al. 1996 for review), including soybean, bean, pea and chickpea

(Akkaya et al. 1992; Akkaya et al. 1995; Maughan et al. 1995; Rongwen et al. 1995; Winter

et al. 1999; Hüttel et al. 1999).

2.4.1.2.5.1 Microsatellite markers in chickpea

SSRs were found to be abundant in the chickpea genome and have moderately high level of

intra-specific polymorphism compared to other marker systems (Sharma et al. 1995b). The

study suggested that SSR markers are well suited for chickpea genome mapping and gene

tagging. As a result, development of SSR markers started in case of chickpea. The first set

of SSR markers in chickpea was developed by Hüttel et al. (1999). Infact, Winter et al.

(1999) reported the first chickpea genetic map based on SSR markers and enlisted 174

primer pairs flanking such loci. Around 280 microsatellite markers were developed at

18

NIPGR (National Institute for Plant Genome Research) using hybridization based

microsatellite enrichment (Sethy et al. 2003, 2006a, b; Choudhary et al. 2006; Bhatia group

unpublished). Lichtenzveig et al. (2005) also developed 233 SSR markers from BAC and

BIBAC libraries of chickpea. In total there are 778 microsatellite markers in chickpea

(Table 1) and they have been extensively used to: (a) estimate genetic diversity among

Cicer species (Udupa et al. 1999; Choumane et al. 2000), (b) construct genetic maps

(Winter et al. 1999; Tekeoglu et al. 2002; Flandez-Galvez et al. 2003; Radhika et al. 2007),

(c) locate QTLs of agronomic importance (Winter et al. 2000; Cho et al. 2002; Rajesh et al.

2002, 2004; Udupa and Baum 2003; Kottapalli et al. 2009).

Table 1: Microsatellite markers available in chickpea

References Number of SSRs developed Hüttel et al. 1999 28 Winter et al. 1999 174 Sethy et al. 2003, 2006a, b, Choudhary et al. 2006, NIPGR, unpublished

280

Lichtenzveig et al. 2005 233 Qadir et al. 2007 63 Total 778

2.4.1.3 Micro-array based molecular markers

Molecular markers based on micro-array based genotyping platform provide the means to

simultaneously screen hundreds to thousands of markers per individual. These markers are

particularly suited for applications related to whole-genome coverage, low costs and large

population sizes. This category includes SNPs, SFPs and DArT markers.

2.4.1.3.1 Single nucleotide polymorphism (SNP)

Polymorphism derived between the individuals/ varieties may arise either due to insertion/

deletion (Indel) of multiple nucleotide bases or single nucleotide substitution. The detection

of variation has led to the development of sequence based molecular marker called SNP

(Wang et al. 1998). In case of plant genomes, SNPs are the most important basic unit of

19

genetic variation and represent commonest class of DNA based genetic markers (Cho et al.

1999; Rafalski 2002). In recent years, these markers have gained considerable importance in

plant genetics and breeding because of their wide distribution, co-dominant inheritance,

high reproducibility and chromosome specific location. The detection and assay of SNPs are

highly amenable to automation (Gupta et al. 2001; Varshney 2008). SNPs are excellent

genetic markers for various applications (Rafalski 2002) including assessment of genetic

diversity (Nasu et al. 2002; Varshney et al. 2007b, 2008; Yang et al. 2004) and evolutionary

studies (Novelli et al. 2004; Carlson et al. 2004, Varshney et al. 2007b), marker-assisted

breeding (Anderson et al. 2005; Van et al. 2008; Varshney et al. 2007a), construction of

high-density genetic map (Cho et al. 1999; Van et al. 2005), detection of genome wide

linkage disequilibrium (Ching et al. 2002; Kim et al. 2006; Mather et al. 2007), population

substructure (Schmid et al. 2006; Caicedo et al. 2007), association mapping of genes

controlling complex traits (Jander et al. 2002; Li et al. 2008) in various plant species.

Due to progress in easy SNP genotyping and assay technologies, these markers are tend be

the most preferred marker system in plant genomic studies in recent years. In chickpea, the

SNPs were detected in coding and genomic regions of chickpea (Rajesh and Muehlbauer

2008; Varshney et al. 2009b) and provide a good source of information for further mapping

and diversity studies in chickpea. Due to the introduction of next generation sequencing

technologies (see Varshney et al. 2009c), SNP identification is expected to be a routine

procedure in case of so called “orphan crops” species including chickpea (Varshney et al.

2009a).

2.4.1.3.2 Diversity array technology (DArT)

Diversity arrays technology (DArT) was developed as a hybridization-based alternative,

which captures the value of the parallel nature of the microarray platform (Jaccoud et al.

20

2001). Subsequently, it was developed for different crops and used in linkage map

construction and diversity analysis. The important plant species for which DArT has been

developed include rice (Xie et al. 2006), barley (Wenzl et al. 2006), Arabidopsis

(Wittenberg et al. 2005), eucalyptus (Lezar et al. 2004), wheat (Semagn et al. 2006b; Akbari

et al. 2006), Cassava (Xia et al. 2005), pigeonpea (Yang et al. 2006) and sorghum (Mace et

al. 2008). DArT simultaneously types several thousand loci in a single assay. DArT

generates whole-genome fingerprints by scoring the presence versus absence of DNA

fragments in genomic representations generated from samples of genomic DNA. DArT

provides high quality markers that can be used for diversity analyses and to construct

medium-density genetic linkage maps. The high number of DArT markers generated in a

single assay not only provides a precise estimate of genetic relationships among genotypes,

but also their even distribution over the genome offers real advantages for a range of

molecular breeding and genomics applications. Of late in chickpea, DArT array with 15,360

features has been developed by ICRISAT in collaboration with DArT Pty Ltd, Australia.

For developing DArT arrays, a set of 96 genotypes representing diverse genotypes from the

reference set and parental genotypes of different mapping populations were used

(unpublished data).

2.4.2 Genetic mapping

Genetic mapping is a procedure of locating the molecular markers or gene loci / QTLs in

order, indicating the relative distances among them, and assigning them to their linkage

groups on the basis of their recombination values from all pairwise combinations. A linkage

map may be thought of as a ‘road map’ of the chromosomes derived from two different

parents (Paterson 1996). The most important use for linkage maps is to identify

chromosomal locations containing genes and QTLs associated with traits of interest; such

maps may then be referred to as ‘QTL’ (or ‘genetic’) maps. ‘QTL mapping’ is based on the

21

principle that genes and markers segregate via chromosome recombination (called crossing-

over) during meiosis (i.e. sexual reproduction), thus allowing their analysis in the progeny

(Paterson 1996). Genes or markers that are close together or tightly-linked will be

transmitted the closer they are situated on a chromosome (conversely, the higher the

frequency of recombination between two markers, the further away they are situated on a

chromosome). Markers that have a recombination frequency of 50% are described as

‘unlinked’ and assumed to be located far apart on the same chromosome or on different

chromosomes (Hartl and Jones 2001; Kearsey and Pooni 1996). Mapping functions are used

to convert recombination fractions into map units called centi- Morgans (cM). Linkage

maps are constructed from the analysis of many segregating markers. The three main steps

of linkage map construction are: (1) production of a mapping population; (2) identification

of polymorphism between parental genotypes for molecular markers and (3) linkage

analysis of markers. Linkage between markers is usually calculated using odds ratios (i.e.

the ratio of linkage versus no linkage). This ratio is more conveniently expressed as the

logarithm of the ratio and is called a logarithm of odds (LOD) value or LOD score (Risch

1992). LOD >3 are typically used to construct linkage maps. A LOD value of 3 between

two markers indicates that linkage is 1000 times more likely (i.e. 1000:1) than no linkage

(null hypothesis). LOD values may be lowered in order to detect a greater level of linkage or

to place additional markers within maps constructed at higher LOD values. Commonly used

software programs include Mapmaker/ EXP (Lander et al. 1987; Lincoln et al. 1993) and

MapManager QTX (Manly et al. 2001), GMendel

(http://cropandsoil.oregonstate.edu/Gmendel), MSTMap (Wu et al. 2008). JoinMap is

another commonly-used program which is generally used for constructing and combining

the genetic maps developed for different mapping populations (Stam 1993). Various

methods used to construct genetic maps are explained in Varshney et al. (2009c) and the

basic principles of linkage map construction are reviewed by Collard et al. (2005).

22

2.4.2.1 Genetic mapping in chickpea

The beginnings of the linkage map development in chickpea were based on morphological

and isozyme loci. However, their small number and the fact that expression of these markers

is often influenced by the environment, makes them unsuitable for routine use. Furthermore,

because of low level of polymorphisms among genotypes of cultivated chickpea (C.

arietinum L.), inter-specific crosses (C. arietinum × C. reticulatum, C. arietinum × C.

echiniospermum) were exploited for genetic mapping (Gaur and Slinkard 1990a,b).

Subsequently, a number of genetic linkage maps of chickpea (Cicer arietinum L.) based on

combination of different marker types have been published. A summary of these maps is

available in Table 2.

2.4.3 Marker-trait association

Marker-trait association can be determined either by linkage-based approach or by linkage-

disequilibrium (LD) based association mapping. In past, in several crops genetic mapping

based approaches were used to identify the QTLs/genes for a trait (Gupta and Varshney

2004) and recently, use of LD-based association mapping also has been used for trait

mapping (Ersoz et al. 2007; Varshney and Tuberosa 2007).

2.4.3.1 Linkage map based marker-trait association

For conducting marker–trait association by using linkage maps, three widely used methods

have been used: single marker analysis (SMA), simple interval mapping (SIM), and

composite interval mapping (CIM) (Tanksley 1993; Liu 1998).

23

Table 2: A list of some genetic maps in chickpea

Mapping population Features of genetic map Genome coverage Reference Inter-specific F2 (C. arietinum × C. reticulatum) 7 linkage groups with 3

morphological and 26 isozymes 200 cM Gaur and Slinkard

1990a, 1990b

F2 (C. arietinum × C. reticulatum) and F2 (C. arietinum × C. echinospermum)

8 linkage groups with 5 morphological and 23 isozymes

257 cM Kazan et al. 1993

F2 (C. arietinum × C. reticulatum) and F2 (C. arietinum × C. echinospermum)

10 linkage groups with 9 morphological, 27 isozyme, 10 RFLP and 45 RAPD loci

527 cM Simon and Muehlbauer 1997

RIL (C. arietinum × C. reticulatum) 11 linkage groups with 120 STMS loci

613 cM Winter et al. 1999

RIL (C. arietinum × C. reticulatum) 16 linkage groups with 118 SSR, 96 DAF, 70 AFLP, 37 ISSR, 17 RAPD, 8 isozyme, 3 cDNA, 2 SCAR and 3 morphological marker

2078 cM Winter et al. 2000

RIL (C. arietinum × C. reticulatum) 9 linkage groups with 89 RAPD, 17 ISSR, 9 isozyme, and 1 morphological marker

982 cM Santra et al. 2000

RIL (C. arietinum × C. reticulatum) 23 linkage groups with RAPD, ISSR and 1 morphological marker

- Hajj- Moussa et al. 2000

RIL (C. arietinum × C. reticulatum) addition of RGA Potkin 1-2 171 to LG5 of Santra et al. (2000)

- Rajesh et al. 2002b

RIL (C. arietinum × C. reticulatum) Extended map of Santra et al. (2000) with 50 SSR and 1 RGA

1,175 cM Tekeoglu et al. 2002

RIL (C. arietinum × C. reticulatum) 12 linkage groups with 47 R gene specific markers integrated to Winter et al. (2000)

2500 cM Pfaff and Kahl 2003

F2 (C. arietinum × C. echinospermum) 570 cM, 8 linkage groups, 14 SSR , 54 RAPD, 9 ISSR, 6 RGA

570 cM Collard et al. 2003

RIL (C. arietinum × C. reticulatum) 10 linkage groups, 16 RAPD, 3 ISSR, 14 STMS

601.2 cM Cobos et al. 2005

RIL (five wide crosses) Consensus map, 8 linkage groups, 555 loci,135 SSRs,33 cross-species, rest others

652.67 cM Millan et al. 2010

Intra-specific RIL (C. arietinum) 14 linkages groups with 68 SSR,

34 RAPD, 4 ISSR, and 5 morphological

297 cM Cho et al. 2002

F2 (C. arietinum) 8 linkage groups with 51 SSR, 3 ISSR, 12 RGA

535 cM Flandez- Galvez et al. 2003

RIL (C. arietinum) 8 linkage groups with 53 STMS 318.2 cM Cho et al. 2004 RIL (C. arietinum) 8 integrated linkage groups with

44 RAPD, 16 ISSR, 165 SSR, 2 RGA, 1 ASAP and two yield related traits

739.6 cM Radhika et al. 2007

F2 (C. arietinum) 10 linkage groups with 82 SSR, 2 EST

724.4 cM Kottapalli et al. 2009

RIL (C. arietinum, fivecrosses) Consensus map, 8 linkage groups, 229 loci, 99 SSR, rest others

426,99 cM Millan et al. 2010

24

2.4.3.1.1 Mapping populations used for QTL interval mapping

The construction of a linkage map requires a segregating plant population (i.e. a population

derived from sexual reproduction). The parents selected for the mapping population will

differ for one or more traits of interest. Population sizes used in preliminary genetic

mapping studies generally range from 50 to 250 individuals (Mohan et al. 1997), however

larger populations are required for high-resolution mapping. Generally in self-pollinating

species, mapping populations originate from parents that are both highly homozygous

(inbred). In cross pollinating species, the situation is more complicated since most of these

species do not tolerate inbreeding. Many cross pollinating plant species are also polyploid

(contain several sets of chromosome pairs). Mapping populations used for mapping cross

pollinating species may be derived from a cross between a heterozygous parent and a

haploid or homozygous parent (Wu et al. 1992). Several different populations may be

utilized for mapping within a given plant species, with each population type possessing

advantages and disadvantages (McCouch and Doerge 1995; Paterson 1996). F2 populations,

derived from F1 hybrids, and backcross (BC) populations, derived by crossing the F1 hybrid

to one of the parents, are the simplest types of mapping populations developed for self

pollinating species. Their main advantages are that they are easy to construct and require

only a short time to produce. Inbreeding from individual F2 plants allows the construction

of recombinant inbred (RI) lines, which consist of a series of homozygous lines, each

containing a unique combination of chromosomal segments from the original parents. The

length of time needed for producing RI populations is the major disadvantage, because

usually six to eight generations are required. Double haploid (DH) populations may be

produced by regenerating plants by the induction of chromosome doubling from pollen

grains, however, the production of DH populations is only possible in species that are

amenable to tissue culture (e.g. cereal species such as rice, barley and wheat). The major

advantages of RI and DH populations are that they produce homozygous or ‘true-breeding’

25

lines that can be multiplied and reproduced without genetic change occurring. This allows

for the conduct of replicated trials across different locations and years. Thus both RI and

DH populations represent ‘eternal’ resources for QTL mapping. Furthermore, seed from

individual RI or DH lines may be transferred between different laboratories for further

linkage analysis and the addition of markers to existing maps, ensuring that all collaborators

examine identical material (Paterson 1996; Young 1996).

2.4.3.1.2 Approaches for QTL interval mapping

Among different statistical analyses for linkage mapping based QTL mapping, SMA (single

marker analysis- also called single-point analysis) is the simplest method for detecting

QTLs associated with single markers. The statistical methods used for SMA include t- tests,

ANOVA, and linear regression. Linear regression is most commonly used because the

coefficient of determination (R2) from the marker explains the phenotypic variation arising

from the QTL linked to the marker. In fact, this method is generally used in BSA approach

for trait mapping. However, the main disadvantages of this method are: (1) the farther a

QTL is from a marker, it is less likely to be detected as the recombination occurring

between the marker and the QTL; (ii) this causes the magnitude of the effect of a QTL to be

underestimated. The use of a large number of segregating markers covering the entire

genome, usually at intervals less than 15 cM, may minimize both problems (Tanksley

1993). Linkage map-based trait mapping approach employs the SIM method that makes use

of linkage maps and analyses intervals between adjacent pairs of linked markers along

chromosomes simultaneously (Lander and Botstein 1989). The use of linked markers for

analyses under SIM is considered statistically more powerful compared to single-point

analysis as the recombination between the markers and the QTL is taken care of Liu (1998).

The CIM approach, however, combines interval mapping with linear regression and

includes additional molecular markers in the statistical model in addition to an adjacent pair

26

of linked markers for interval mapping (Jansen and Stam 1994). This method is more

precise and effective at mapping QTLs as compared to single-point analysis (SMA) and

SIM, especially when linked QTLs are involved.

2.4.3.1.3 QTL analysis for drought tolerance

Drought tolerance is the complex phenomenon involving many known and unknown

pathways. The dissection of QTLs for such a complex trait is challenging and many

attempts have been made in different crop species. In order to find the QTLs for drought

tolerance, the traits like stomatal conductance (Sanguineti et al.1999; Juenger et al. 2005;

Price et al. 1997), transpiration efficiency (Juenger et al. 2005; Specht et al. 2001; Kholova

et al. 2010), osmotic adjustment (Diab et al. 2004; Saranga et al. 2004;Robin et al. 2003),

relative water content (Diab et al. 2004; Sanguineti et al.1999), canopy temperature

(Saranga et al. 2004), drought sensitivity index (Sanguineti et al.1999); leaf ABA (Kholova

et al. 2010); chlorophyll content (Shen et al. 2007); water use efficiency (Quarrie et al.

2005), root traits (Chandra Babu et al. 2003; Li et al. 2005) and some yield related traits

(Diab et al. 2004; Saranga et al. 2004; Moreau et al. 2004; Xu et al. 2005; Xiao et al. 2005;

Specht et al. 2001; Quarrie et al. 2005; Dashti et al. 2007). Studies at ICRISAT had revealed

the importance of root traits in drought tolerance in chickpea.

2.4.3.1.4 Importance of root traits in chickpea drought tolerance

The yields chickpea genotypes under rainfed and irrigated conditions were compared at

ICRISAT in order to compare the yields under drought conditions and the potential yields

(Saxena 2003). The study indicated that genotype ICC 4958 exhibited the best performance

not only in field trials in ICRISAT but also in other Mediterranean environments, which had

higher root biomass. Subsequent studies at ICRISAT on 12 diverse chickpea genotypes

showed that the prolific root system upto the depth of 15-30 cm contributed positively

27

towards the seed yield under moderate to severe terminal drought conditions (Kashiwagi et

al. 2006). The advantage of deep root systems towards drought tolerance mechanism was

well substantiated in soybean (Kaspar et al. 1978), common beans (Sponchiado et al. 1989)

and chickpea (Silim and Saxena 1993). Some major root attributes such as greater efficiency

in water absorption per unit root length density, ability to change the rooting pattern across

the soil depths for efficient access of soil moisture and the ability to produce larger root

surface area per unit root biomass make chickpea the best choice for dryland cropping

systems compared to other legumes and cereals (Thomas et al. 1995, Benjamin and Nielsen

2006). Chickpeas also found to have higher root surface to root weight ratio compared to

soybean and field peas (Benjamin and Nielsen 2006). These results suggest that chickpeas

are better equipped towards tolerance to drought stress and further improvement of root

traits would be one of the promising approaches to improve drought avoidance in chickpea

under terminal drought environments (Gaur et al. 2008).

2.4.3.1.5 QTL analysis for drought tolerance related traits

Compared to the conventional breeding approaches for improved productivity under water

limited environments, genomics offers great opportunities for dissecting quantitative traits

into their single genetic determinants (Young 1996; Dudley 1993; Tanksley 1993; Lee

1995; Beavis and Kein 1996; Quarrie 1996; Prioul et al. 1997; Tuberosa et al. 2002).

Identification of QTLs is paving the way to MAS (Ribaut et al. 2002; Morgante and

Salamini 2003) and assisted pyramiding of the beneficial QTL alleles. Marker-assisted

breeding reduces the effect of environmental conditions during the selection process, which

is a major hindrance in conventional breeding under drought. The increasing number of

studies reporting QTLs for drought tolerance related traits in different crops under drought

stress (Table 3) indicates a growing interest in this approach. With the invention of other

genomic tools, sequencing and bioinformatics, new dimensions for deciphering and

28

manipulating the genetic basis of drought tolerance can be achieved (Tuberosa et al. 2002;

Varshney et al. 2005b; Tuberosa et al. 2005).

Although considerable progress has been made in identifying QTLs in chickpea related to

Fusarium wilt and Ascochyta blight disease resistance (Table 4), information on the genetic

basis of traits related to drought tolerance in chickpea is limited. A deep root system capable

of extracting additional soil moisture should positively impact yield under drought stress

environments.

29

Table 3: Some QTL studies related to drought tolerance in selected crop species

Crop Cross Traits studied Reference Barley Tadmor × Er/Apm Osmotic adjustment, leaf relative

water content, grain yield Diab et al. 2004

Cotton G. hirsutum × G. barbadense

Canopy temperature, osmotic potential, dry matter, seed yield

Saranga et al. 2004

Maize F2 × F252 Silking date, grain yield, yield stability

Moreau et al. 2004

Maize Os420 × IABO78 Stomatal conductance, drought sensitivity index, leaf temperature, leaf relative water content, anthesis-silking interval, grain yield

Sanguineti et al. 1999

Maize X178 × B73 Yield traits measured in drought conditions-grain yield, 100 kernel weight, number of ears per plant

Xiao et al. 2005

Pearl millet PRLT 2/89-33 × H 77/833-2 Leaf ABA, transpiration efficiency, transpiration at vapour pressure deficit

Kholova et al. 2010

Pearl millet 863B-P2 × ICMB 841-P3 Leaf ABA, transpiration efficiency, transpiration at vapour pressure deficit

Kholova et al. 2010

Rice Zhenshan97B × Milyang 46 Chlorophyll content Shen et al. 2007 Rice CT9993 × IR62266 Root morphology, plant height,

grain yield Chandra Babu et al. 2003

Rice IR62266 × IR60080. Osmotic adjustment Robin et al. 2003 Rice IRAT109 × Yuefu Root traits Li et al. 2005 Rice Teqing × Lemont Phenology, yield components Xu et al. 2005 Rice Nipponbare × Kasalath Stomatal frequency, leaf rolling Ishimaru et al. 2001 Rice Azucena × Bala Stomatal conductance, leaf rolling,

heading date Price et al. 1997

Sorghum B35 × Tx7000 Stay green, chlorophyll content Xu et al. 2000 Soybean Minsoy × Noir 1 Carbon isotope descrimination

(CID), Transpiration efficiency, yield components

Specht et al. 2001

Wheat SQ1 × Chinese spring Water use efficiency, grain yield Quarrie et al. 2005 Wheat Beaver × Soissons Flag leaf senescence Verma et al. 2004 Wheat Double haploid lines

Chinese spring × SQ1 Stress susceptibilty index (SSI), mean productivity index (MP), tolerance index, stress tolerance index (STI) and yield components

Dashti et al. 2007

30

Table 4: QTL studies related to agronomic traits in chickpea

Cross Type Traits Reference ICCV96029 × CDC Frontier Desi × Kabuli Ascochyta blight Tar’an et al. 2007 ILC72 × Cr5-10 C. arietinum × C. reticulatum Ascochyta blight Cobos et al. 2006 ILC3279 × WR315 Kabuli × Desi Ascochyta blight Iruela et al. 2006 Hadas × ICC 5810 Kabuli × Desi Ascochyta blight,

time of flowering Lichtenzveig et al. 2006

PI359075 × FLIP84-92C C. arietinum Ascochyta blight Cho et al. 2004 FLIP84-29C × PI599072 C. arietinum × C. reticulatum Ascochyta blight Tekeoglu et al. 2002 Lasseter × PI527930 C. arietinum × C. echinospermum Ascochyta blight Collard et al. 2003 ICC12004 × Lasseter Desi Ascochyta blight Flandez-Galvez et al. 2003 ILC3279 × CA2156 C. arietinum Ascochyta blight Millan et al. 2003 ICC4958 × PI489777 C. arietinum × C. reticulatum Ascochyta blight Rakshit et al. 2003 FLIP84-92C × PI599072 C. arietinum × C. reticulatum Ascochyta blight Rakshit et al. 2003 ILC1272 × ILC3279 Kabuli Ascochyta blight Udupa and Baum 2003 FLIP84-92C × PI599072 C. arietinum × C. reticulatum Ascochyta blight Santra et al. 2000 PI359075 × FLIP84-92C(2) C. arietinum Ascochyta blight Tekeoglu et al. 2000a Blanco Lechoso × Dwelley C. arietinum Ascochyta blight Tekeoglu et al. 2000a FLIP84-92C × PI599072(3) C. arietinum × C. reticulatum Ascochyta blight Tekeoglu et al. 2000a CDC Frontier × ICCV 96029 C. arietinum Ascochyta blight Anbessa et al. 2009 CA2156 × JG62 Kabuli × Desi Fusarium wilt Cobos et al. 2005 CA2139 × JG62 Kabuli × Desi Fusarium wilt Cobos et al. 2006 C104 × WR315 C. arietinum Fusarium wilt Sharma et al. 2004 ICC4958 × PI489777 C. arietinum × C. reticulatum Fusarium wilt Tekeoglu et al. 2000b ICC4958 × PI498777 C. arietinum × C. reticulatum Fusarium wilt Winter et al. 2000 C-104 × WR-315 C. arietinum Fusarium wilt Tullu et al. 1998 C104 × WR315 C. arietinum Fusarium wilt Mayer et al. 1997 FLIP84-92C × PI599072 C. arietinum × C. reticulatum Beet armyworm Clement et al. 2010

31

A preliminary study carried out at ICRISAT on drought tolerance using single marker linear

regression analysis in chickpea (Annigeri × ICC 4958) indicated SSR marker “TAA170”

associated with the drought tolerance related traits like root length, root length density and

shoot dry weight. Apart, two mapping populations segregating for drought tolerance related

traits- ICC 4958 × ICC 1882 and ICC 283 × ICC 8261- were developed based on the

physiological studies (Kashiwagi et al. 2005, 2006), to dissect the QTLs for drought related

root traits.

After identifying important QTLs, the next step involves the identification of candidate

sequences, validate their role and proceed with the direct manipulation using the gene itself

as marker for MAS (Tuberosa and Coraggio 2004). The recent progress in the profiling of

transcriptome, proteome and metablome offers the possibility of investigating the response

of genes to drought and other stresses. Current focus in chickpea functional genomics

should be able to coordinate resources around the world and take full advantage of

functional genomics for chickpea improvement (Coram and Pang 2007). Further

development with respect to molecular breeding for agronomically important traits could lie

in the integration of aspects of physiology and biotechnology towards plant breeding (Blum

and Nguyen 2004). The characterization of key plant physiological mechanisms that restrain

performance under drought, together with the associated regulatory genes, could therefore,

facilitate the development of improved crop varieties showing water use efficiency and

drought tolerance.

2.4.3.2 Association mapping

Association mapping, also known as linkage disequilibrium (LD) mapping, has emerged as

a tool to resolve complex trait variation down to the sequence level by exploring historical

and evolutionary recombination events at the population level (Nordborg and Tavare 2002;

32

Risch and Merikangas 1996). As a new alternative to traditional linkage analysis,

association mapping offers three advantages, (i) increased mapping resolution, (ii) reduced

research time, and (iii) greater allele number (Yu and Buckler 2006). Since its introduction

to plants (Thornsberry et al. 2001), association mapping has continued to gain favourability

in genetic research because of advances in high throughput genomic technologies, interests

in identifying novel and superior alleles, and improvements in statistical methods (Zhu et al.

2008).

2.4.3.2.1 Strategies for association analysis

While using the association or LD mapping approach, the statistical power of associations is

determined by the extent of LD with the causative polymorphism, as well as sample size

used for the study (Wang and Rannala 2005). The decay of LD over physical distance in the

study population determines the marker density required and the level of resolution that may

be obtained in an association study. The most commonly used summary statistic for

estimation of LD within the association study framework is known as r2 (Hill and Robertson

1968; Lewontin 1988). The r is the Pearson’s (product moment) correlation coefficient of

the correlation that describes the predictive value of the allelic state at one polymorphic

locus on the allelic state at another polymorphic locus, where r2 is the squared value of

correlation coefficient that is also called coefficient of determination and it explains the

proportion of a sample variance of a response variable that is explained by the predictor

variables when a linear regression is performed (Ersoz et al. 2007). Lewontin’s D is another

summary statistic for LD that is commonly used and describes the difference between the

coupling gamete frequencies and repulsion gamete frequencies at two loci. From D, a

second measure of LD, that is, normalized D´ can also be estimated. It is important to

estimate the rate of decay of LD with physical distance, to be able to extrapolate

information gathered from a small collection of sampled loci to the whole genome

33

investigated. This extrapolation is essential for association mapping study design, since it

may be used for determining the marker density required for scanning previously

unexplored regions of the genome, as well as determining the maximum resolution that can

be achieved for genotype–phenotype associations for the study population. Another

important constraint for the use of association mapping for crop plants is unidentified

population sub-structuring and admixture due to factors such as adaptation or domestication

(Wright and Gaut 2005). Population structure creates genome-wide LD between unlinked

loci. When the allele frequencies between sub-populations of a species are significantly

different, due to factors such as genetic drift, domestication, or background selection,

genetic loci that do not have any affect whatsoever on the trait may demonstrate statistical

significance for their co-segregations with a trait of interest. In cases where the population

structuring is mostly due to population stratification (Bamshad et al. 2004; Pritchard 2001),

three methods